ClawHub
Back to skill

Security audit

Agentic Beehive MCP Server

Security checks across malware telemetry and agentic risk

Overview

This is a coherent local MCP server for agent state and external endpoint tracking, with disclosed persistence and network polling risks that users should manage.

Install only if you want a local MCP server that stores agent state and can contact registered external endpoints. Pin or review the fastmcp dependency, register only trusted endpoints, avoid storing secrets in metadata or alerts, and periodically review or clear the local beehive.db state.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding
The skill advertises installation and registration of a Python-based MCP server and explicitly installs dependencies, which implies executable code with network and environment access, yet it declares no permissions or capability boundaries. This creates a transparency and trust problem: operators may enable the skill without understanding that it can access external systems or local environment data, increasing the risk of over-privileged deployment or unintended data exposure.

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
`colony_poll` performs outbound HTTP GET requests to endpoints that can be registered via `colony_register`, with no allowlist, scheme restriction, hostname validation, or consent gate. In an agent/server context this creates an SSRF-style primitive and broader-than-necessary network reachability for a scheduling/status hub, allowing probing of internal services or unintended egress.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The HTTP polling occurs silently from the tool invocation path and the user-facing interface does not prominently disclose that calling `colony_poll` will send network traffic to the stored endpoint. In agent environments, undisclosed outbound requests can surprise operators, leak metadata such as IP/user-agent/timing, and amplify the SSRF/evasion risk introduced by arbitrary endpoint registration.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.